Preparation and characterization of MgG rods
Galvanic cells generally consist of two kinds of metals with different electrode potentials32,33. To engineer MgG, Mg rods were placed in a neutral solution containing PtCl62− at room temperature, and then the Mg galvanic cells (Mg2+/Mg: −2.372 V; PtCl62−/PtCl42−: 0.726 V; PtCl42−/Pt: 0.758 V) were successfully constructed by in situ reducing Pt on the surface of Mg rods (Fig. 2a). It is known that the ion concentration (PtCl62−) and the immersion time directly affect the Pt loading component via the self-assembly process, which may have an impact on the efficiency of H2 generation. To optimize the MgG parameters, the H2 generation performance of MgG rods prepared under different immersion time and concentrations of PtCl62− was determined using gas chromatography. The H2 generation performance was enhanced with increasing ion concentrations and immersion time, and then reached a steady-state (Supplementary Figs. 1, 2). After MgG (PtCl62− concentration: 0.3%, immersion time: 1 min) construction, scanning electron microscopy (SEM) images and energy dispersive spectrometry (EDS) elemental mapping of a MgG rod showed homogeneous distribution of Mg and Pt elements (Fig. 2b and Supplementary Figs. 3, 4). Transmission electron microscopy (TEM) image and elemental mapping showed that the Pt nanoparticles (NPs, ~3 nm) were uniformly distributed over the Mg rods (Supplementary Fig. 5). XRD analysis also showed the peaks of Pt (JCPDS No. 04-0802) emerging from the as-prepared MgG rods, suggesting that Pt NPs were in situ reduced on the surface of the Mg rods (Fig. 2c and Supplementary Fig. 6).
a Schematic illustration of the synthesis process of Mg galvanic cells and the H2 production mechanism. b The representative SEM image and EDS element mapping of MgG rods from three independent samples. c The XRD patterns of MgG rods. d Photograph showing H2 generation in PBS from Mg and MgG. e Schematic illustration of H2 generation detected by the MB probe. f Time-dependent absorption spectra of the MB solution (pH = 6.5) with MgG rods added. g Comparison of MB reduction by Mg rods and MgG rods. h Time-dependent H2 generation measured by gas chromatography from Mg or MgG rods in PBS (pH = 6.5, n = 3 biologically independent samples). i H2 generation profiles of MgG rods in PBS solutions with different pH values measured by gas chromatography. j The Tafel curve of Mg rods and MgG rods. k The corrosion potentials of Mg rods and MgG rods (n = 5 biologically independent samples). Data are presented as mean values ± SD.
For Mg-Pt galvanic cells, once the metal electrode was placed into the aqueous environment, a spontaneous redox reaction occurred. Electrons (e−) would flow out of the negative electrode (Mg electrode, Mg-2e− = Mg2+) and flow into the positive electrode (Pt electrode, 2H2O + 2e− = 2OH− + H2), leading to the water etching of Mg and the generation of H2. Indeed, by immersion in water, MgG rods could rapidly react with H2O and generate lots of H2 gas bubbles, while bare Mg rods virtually generated no obvious bubbles (Fig. 2d). To further verify H2 generation, methylene blue (MB) was used as the probe to detect H2 produced from MgG. The blue-colored MB could be quickly reduced into colorless MBH2 by H2 (Fig. 2e)34. Notably, the MB characteristic peak in the MgG group was significantly reduced, while that in the bare Mg rod group only showed a slight decrease, indicating that MgG rods had a competent ability to generate H2 gas by reacting with water (Fig. 2f, g and Supplementary Fig. 7). Quantitative measurement of H2 generation from the MgG and Mg rods was further conducted by gas chromatography. It was found that compared to bare Mg rods, MgG rods were able to generate more H2 gas, continuously for more than 1 week (Fig. 2h). In addition, the H2 production rate of MgG rods increased under lower pH, indicating that the MgG rods would show better H2 generation capacity in the weak acid TME (Fig. 2i and Supplementary Fig. 8). During the H2 generation process, no significant O2 was generated because Mg was oxidized in the negative electrode reaction (Supplementary Fig. 9). In addition, the XPS spectra of MgG after the H2 generation process showed strong signals from Mg (II), further proving that MgG rods were sacrificed during the generation of H2 (Supplementary Fig. 10). Essentially, the galvanic cell is an electrochemical etching reaction, and a lower corrosion potential indicates an easier occurrence of oxidation corrosion reaction30,35. Based on the Tafel curves, the corrosion potential of MgG was much lower than that of Mg, explaining the reason that MgG would react with water more efficiently (Fig. 2j, k).
In vitro hydrogen therapy with MgG rods
With continuous H2 generation ability, we expected the utilization of MgG rods for further H2 gas therapy (Fig. 3a). It has been reported that a high concentration of H2 kills cancer cells by inhibiting cell mitochondrial respiration16. Therefore, the effect of MgG rods on inhibiting mitochondrial respiration was evaluated by measuring the membrane potential changes, with JC-1 dye as the fluorescent indicator. JC-1 dye forms red fluorescent aggregates in the normal mitochondrial membrane, while green fluorescent monomers exist in the damaged mitochondrial membrane36. Weak green fluorescence and strong red fluorescence were observed in the control group, indicating intact mitochondria of those cells (Fig. 3b and Supplementary Fig. 11). While cells treated with Mg rods showed only a slight green fluorescence increase, the strongest green fluorescence was observed in cells after being treated with MgG rods, indicating that the continuous and abundant H2 generation would significantly affect the cellular mitochondrial respiration. Moreover, rather weak green fluorescence occurred in the Pt and Mg(OH)2– treated groups, demonstrating no significant mitochondrial damage caused by Pt or Mg(OH)2 alone. On the other hand, while the equilibrium of redox homeostasis inside cells is important for their survival37,38, it is known that breaking the intracellular redox homeostasis equilibrium would affect cell growth39,40. As expected, H2 generated from MgG rods drastically altered the oxidative stress state of cancer cells with time (Fig. 3c, d and Supplementary Fig. 12). As a direct energy source for basic life activities, adenosine triphosphate (ATP) affects the living state of cancer cells. It was found that the ATP concentration within 4T1 cells showed a remarkable decrease over time after being treated with MgG rods (Fig. 3e), suggesting that the cellular activity was inhibited by the continuously generated H2, which thus caused a remarkable reduction in cellular energy. The above results demonstrated that the continuous H2 generation by Mg-based micro-galvanic cells would cause effective mitochondrial dysfunction and intracellular redox homeostasis destruction.
a Schematic illustration showing the mechanims of hydrogen therapy with MgG for cancer cell killing. b Detection of mitochondria membrane potentials by confocal fluorescence images of 4T1 cells stained with JC-1 dye. c Time-dependent changes of intracellular ROS in 4T1 cells during different treatments (n = 5 biologically independent samples). d Flow cytometry data to show DCF-positive 4T1 cells after different treatments. e Time-dependent changes of intracellular ATP contents in 4T1 cells during different treatments (n = 5 biologically independent samples). f Relative viabilities of 4T1 and CT26 cells after various treatments (control, Mg, Pt, Mg(OH)2, and MgG, n = 5 biologically independent samples). g Flow cytometry analysis of 4T1 cells after various treatments using an Annexin V-FITC/PI kit. A representative image of three biologically independent samples from each group is shown in (b). Data are presented as mean values ± SD.
To test their in vitro cell-killing efficacy, MgG rods were incubated with 4T1 murine breast cancer cells or CT26 colon adenocarcinoma cells, with bare Mg rods used as the control. There were only slight changes in cell viabilities after treatment with Mg rods, while most cells were damaged after being treated with MgG rods, indicating that the sustained H2 generation at a high level could significantly lead to cell death (Fig. 3f). Notably, Pt NPs, Mg(OH)2, and Mg2+ exhibited only slight toxicity (Fig. 3f and Supplementary Fig. 13). Since alloying may increase the corrosion rate of Mg, the H2 generation performance and cell-king effect of MgG rods and other commercialized Mg alloys (Mg, MgZnCa, MgAl) were evaluated. The MgG rods exhibited the strongest cell-killing effect, probably due to their superior H2 generation capacity (Supplementary Figs. 14, 15). To further verify the ability of H2 to kill cells, 4T1 and CT26 cells were treated with pure H2 released from a hydrogen balloon. It was found that a high concentration of H2 could indeed inhibit the proliferation of cancer cells (Supplementary Fig. 16). Moreover, the flow cytometry analysis further confirmed that MgG treatment caused significant apoptosis of 4T1 and CT26 cells (Fig. 3g and Supplementary Fig. 17). All of the above results demonstrated that the continuous H2 generation originating from MgG rods could lead to the effective killing of cancer cells by inhibiting mitochondrial respiration and disrupting redox homeostasis.
In vivo hydrogen therapy with MgG rods
Next, we wondered whether such MgG rods could be used for in vivo antitumor therapy. First, ultrasonic imaging was performed to monitor hydrogen gas generation in the tumor, as gas bubbles would offer strong contrast under ultrasound imaging. After MgG rods were implanted into the tumor, strong ultrasound signals appeared in the tumor for more than 48 h, suggesting efficient gas production (Fig. 4a, b and Supplementary Fig. 18). In contrast, the ultrasound signals in Mg rod implanted tumors were much weaker, indicating the insignificant H2 generation from the bare Mg rods (Fig. 4b and Supplementary Fig. 19). In addition, SEM images and EDS element mapping of MgG rods after implantation into the tumor for 4 and 24 h showed that the galvanic cell structure accelerated the corrosion of Mg to generate enough H2. Despite the obvious corrosion of MgG rods, Pt NPs remained on the surface of those rods at 24 h, allowing further in vivo corrosion of MgG and thus the generation of H2 for gas therapy (Supplementary Fig. 20). Notably, the MgG rods implanted into the tumor indeed showed efficient H2 generation at 4 and 24 h as evidenced by gas chromatography (Supplementary Fig. 21). After etching MgG, we expected that the residual product Mg(OH)2 as an alkaline substance would be able to neutralize the acidic tumor pH. Therefore, the time-dependent pH change curves of tumors in mice post intratumoral implantation with MgG rods were monitored by a pH microsensor (Fig. 4c). The tumor pH was rapidly neutralized to the neutral level and could be maintained for up to 72 h. Furthermore, ex vivo fluorescence imaging of tumors after MgG rods implantation was performed using a pH-responsive fluorescent probe (Fig. 4d, e). Consistently, the tumor pH reached the peak level at 24 h post-MgG implantation, and such neutralization effect could be maintained for 72 h. Notably, although the metal-water reaction is exothermic, there was no temperature change at the tumor site after intratumoral implantation of MgG rods, suggesting that the reaction of MgG with water was moderate and would not induce a significant thermal effect (Supplementary Fig. 22). Solid tumors are characterized by an acidic microenvironment, which may impede effective antitumor T-cell immune responses. Specifically, CD8+ T cells tend to become anergic and apoptotic when exposed to a low pH environment. The neutralization of tumor acidity would be favorable for antitumor immune responses41,42. Therefore, MgG-induced antitumor immune responses were evaluated on 4T1 tumor-bearing mice. These mice post-implantation of MgG at day 0 were sacrificed at day 6, and their tumors were collected to determine the levels of immune cells. Compared with the other groups, the Mg(OH)2 group showed a slight regulatory effect, while the tumors from the MgG group exhibited reduced populations of MDSCs and an increased percentage of total T cells (especially CD8+ T cells). MDSCs, as immunosuppressive cells, could promote tumor progression by inhibiting antitumor immunity, while T cells (especially CD8+ T cells) would ensure effective tumor cell killing. Therefore, MgG implantation would thus modulate the immunosuppressive TME for enhanced tumor therapy (Fig. 4f–i, and Supplementary Figs. 23, 24).
a In vivo time-dependent ultrasonic imaging of 4T1 tumor-bearing mice after intratumoral implantation with MgG rods. b Quantitative analysis of signal intensities based on ultrasonic imaging data. c Time-dependent pH value changes of tumors in mice post intratumoral implantation with Mg rods or MgG rods. d Ex vivo fluorescence images of tumors after implantation of MgG for different periods of time. e Tumor fluorescence (FL) signal ratios (480/440) based on ex vivo fluorescence imaging data in (C). f–h The quantification results of MDSCs (CD45+CD11b+Gr-1+, f), T cells (CD3+, g), CD8+ T cells (CD3+CD8+, h) by flow cytometry on day 6 post-MgG implantation. i The flow cytometric analysis results of CD8+ T cells (CD3+CD8+) within the tumors after different treatments. j The growth curves of tumors after various treatments. k Survival rates of tumor-bearing mice after various treatments. l Microscopy images of H&E and TUNEL stained tumor slices collected from mice post different treatment groups. m In vivo bioluminescence images of mice bearing subcutaneous 4T1 tumors expressing firefly luciferase (Luc-4T1) to display the therapeutic efficacy of mice after various treatments. n = 3 biologically independent animals in (b, c). n = 4 biologically independent animals in (e). n = 5 biologically independent animals in (f–h). n = 10 biologically independent animals in (j). A representative image of three biologically independent animals from each group is shown in (a, l). Data are presented as mean values ± SD. P values calculated by the two-tailed student’s t test are indicated in the Figures.
Next, we carefully investigated the antitumor efficacy of H2 gas therapy based on the micro-galvanic cell strategy. Mice bearing subcutaneous 4T1 tumors were randomly divided into seven groups (n = 10) when their tumor volumes reached ~150 mm3: (I) Control; (II) Mg rods implantation; (III) Pt NPs (remaining after dissolving MgG rods); (IV) Mg(OH)2 (remaining after the MgG rods produced H2); (V) MgG rods implantation; (VI) Mg rods implantation (three times, 0th day, 4th day, and 8th day); and (VII) MgG rods implantation (three times, 0th day, 4th day, and 8th day). For Mg or MgG rods implantation, two Mg or MgG rods (D = 0.5 mm, L = 4 mm) were implanted into each tumor (∼7 mm × 7 mm) (Supplementary Fig. 18). Pt NPs- and Mg-implanted groups had no significant tumor-suppressing effect (Fig. 4j, k). The tumors of the Mg(OH)2 group showed slightly delayed growth, likely due to the neutralization of tumor pH. Remarkably, the tumor growth of the MgG rods implantation group was significantly inhibited, especially for those with three times of MgG rod implantations. Notably, MgG implantation could significantly prolong animal survival, and four out of ten mice survived for more than 60 days after three times of MgG-implantations (Fig. 4k and Supplementary Fig. 26). Hematoxylin and eosin (H&E) staining and TdT-mediated dUTP nick-end labeling (TUNEL) staining of tumor slices further confirmed that MgG rods implantation could induce severe tumor cell apoptosis (Fig. 4l and Supplementary Fig. 25). In addition, there was no significant body weight loss for mice implanted with MgG rods, indicating no obvious side effects induced by MgG-based H2 therapy (Supplementary Fig. 27). Moreover, mice bearing subcutaneous 4T1 tumors expressing firefly luciferase (Luc-4T1) were also used to intuitively display the therapeutic efficacy. MgG rods implantation group showed the weakest bioluminescence signal, and demonstrated the best tumor growth inhibition (Fig. 4m).
In vivo hydrogen therapy of different tumor models vith MgG rods
Encouraged by the superior therapeutic performance of MgG rods for the treatment of 4T1 tumors, we then evaluated the therapeutic efficacy of MgG-induced H2 therapy on other tumor models. CT26 tumor-bearing mice were randomly assigned into five groups (n = 10) when their tumor volume reached ~150 mm3 (Fig. 5a). Following a similar trend to the results of the 4T1 model, Mg rods implantation was essentially unable to inhibit tumor growth, even after three times of implantation. As expected, MgG rods implantation significantly inhibited the tumor growth, with ~40% of tumors being completely eliminated after three times of MgG implantations (Fig. 5b, c and Supplementary Figs. 28, 29). Tumor cells were severely damaged by MgG rods implantation while the other groups showed no obvious cell damage as determined by H&E and TUNEL staining (Fig. 5d and Supplementary Fig. 25). In addition, there was no significant body weight loss for the mice implanted with MgG rods, indicating no obvious side effects induced by MgG-based H2 therapy (Supplementary Fig. 30). Therefore, the MgG-induced H2 therapy strategy on CT26 model also showed great therapeutic effects.
a Scheme of the subcutaneous CT26 mouse tumor model for MgG-based hydrogen therapy. b, c Tumor growth curves (b) and survival rates (c) of CT26 tumor-bearing mice after different treatments as indicated. d Microscopy images of H&E and TUNEL stained CT26 tumor slices collected from different groups. e Scheme of the PDX tumor model in mice for MgG-based hydrogen therapy. f, g Tumor growth curves (f) and survival rates (g) of nude mice after different treatments. h Microscopy images of H&E and TUNEL stained PDX tumor slices collected from different groups. i Scheme of the VX2 liver rabbit tumor model for MgG-based hydrogen therapy. j, k Tumor growth curves (j) and survival rates (k) of rabbits after different treatments. l Microscopy images of H&E and TUNEL stained VX2 tumor slices collected from different groups. n = 10 biologically independent animals in (b). n = 7 biologically independent animals in (f). n = 6 biologically independent animals in (j). A representative image of three biologically independent animals from each group is shown in (d, h, l). Data are presented as mean values ± SD.
Next, we further tested our MgG-induced H2 therapy strategy to treat a more clinically relevant tumor model. A cervical PDX tumor model43,44, in which tumors from real human patients are implanted into immunodeficient mice, was further used to evaluate the efficacy of MgG-induced H2 therapy. To establish a PDX tumor model, tumor tissues were surgically resected from a cervical cancer patient and cut into pieces, which were implanted into nude mice (Fig. 5e). For MgG-induced H2 therapy, PDX tumor-bearing nude mice were randomly assigned into five groups (n = 7) when the tumor volume reached ~150 mm3, and treated via the same parameters as that used for the treatment of the CT26 tumor model. Although it was difficult to completely eliminate those PDX tumors, MgG rod implantation still significantly extended the survival time of mice compared with that of the other control groups (Fig. 5f, g and Supplementary Fig. 31), and caused obvious tumor damage as revealed by H&E and TUNEL staining of tumor slices (Fig. 5h and Supplementary Fig. 25). In addition, there was no significant decrease of body weight for the mice implanted with MgG rods, indicating good safety of MgG-based H2 therapy (Supplementary Fig. 32).
In addition to the mouse model experiment, the therapeutic efficacy of our strategy in the treatment of a larger animal tumor model was also evaluated. Rabbit-bearing VX2 tumors were randomly allocated and treated by intratumoral implantation of MgG rods (Fig. 5i). Due to the large rabbit tumor sizes (∼800 mm3, five times larger than that on mice), three Mg or MgG rods (L = 8.0 mm, D = 0.8 mm) were implanted into each tumor (∼12 mm × 12 mm) for rabbit treatment (Supplementary Fig. 18). MgG-implanted tumors were effectively inhibited by continuous H2, while tumors of the control group or Mg rods implantation group showed rapid growth (Fig. 5j, k and Supplementary Fig. 33). In fact, tumors on two out of six rabbits were completely eliminated without re-growth and survived for over 120 days after H2 treatment with three implantation cycles (Fig. 5k). Moreover, the tumor slices collected from rabbits treated with MgG rods also showed severe histological damage and a high level of cell apoptosis compared to the other treatment groups (Fig. 5l and Supplementary Fig. 25). All these results showed that VX2 tumors with larger sizes could also be effectively inhibited after MgG rods implantation, illustrating the capability of MgG-based H2 therapy to kill large-sized tumors.
In vivo safety and biodegradation behaviors of MgG rods
Last but not the least, it is important to evaluate the safety of MgG rods before future clinical trials. 4T1 tumor-bearing mice were intratumorally implanted with MgG rods for different time points. First, the Mg levels in the tumors of mice after being implanted with MgG rods for different times were measured by inductively coupled plasma optical emission spectrometry (ICP-OES) (Fig. 6a). The Mg content in the tumor gradually decreased over time, and most of the MgG rods were degraded within 15 days. Such degradable behavior of MgG rods was mainly due to the galvanic-cell-accelerated water etching of Mg after implantation. The formed Mg(OH)2 would further react with the H+ in the TME to form Mg2+ ions. Moreover, there was no significant increase of Mg levels in the blood and main organs (liver, spleen, kidney, heart, and lung) after MgG rods were implanted into the body, indicating that Mg2+ generated from the degraded MgG would not affect the equilibrium of Mg levels in the body (Fig. 6b). In addition, the biodistribution of Pt NPs after degradation from MgG in the tumor and the main organs was studied. Pt content in the tumor was gradually decreased over time, and the major distribution of Pt was detected in the kidney, indicating the renal clearance of those Pt NPs with their ultrasmall sizes (Fig. 6c, d). Meanwhile, no obvious histological damage was observed on organ slices from the MgG rods treated mice (Fig. 6e). Other main indicators were further evaluated as to whether rapid degradation would result in potentially toxic side effects. Alanine aminotransferase (ALT), aspartic aminotransferase (AST), and alkaline phosphatase (ALP), all of the important liver function markers, were within the reference range comparable to the control group, indicating that MgG rods implanted caused no significant hepatotoxicity (Fig. 6f). As an indicator of kidney function, the blood urea levels in the blood of treated mice were also within the normal range (Fig. 6f). For the hematological assessment, the white blood cells, red blood cells, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration, and platelets were selected (Supplementary Fig. 34), and all these hematological assay data were found to be normal in the MgG-treated groups compared with that in the control group. Moreover, there was no obvious pH value variation in blood of Balb/c mice after being implanted with MgG rods (Fig. 6g). Taken together, all the above results illustrated that MgG rods could be rapidly degraded after implantation and displayed no obvious toxicity to the treated animals.
a The Mg levels in tumors of mice after being implanted with MgG rods for different periods of time. b The Mg levels in major organs and blood of mice after being implanted with MgG rods for different periods of time. The Mg contents were measured by ICP-OES. No notable variation of Mg levels was found in the blood or organs of implanted mice compared to the control. c The Pt levels in tumors of mice after being implanted with MgG rods for different periods of time (1, 2, 3, 7, and 15 D). d The Pt levels in major organs of mice after being implanted with MgG rods for different periods of time (1, 2, 3, 7, and 15 D). e Hematoxylin and eosin (H&E) staining of mouse major organs (liver, spleen, kidney, heart, and lung) to examine the histological changes after implantation of MgG rods in mice. f Blood biochemistry data of Balb/c mice after being implanted with MgG rods for different periods of time (0, 6 h, 24 h, 48 h, 72 h, 7 days, and 15 days). The measured indexes included alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and blood urea. g Blood pH value variation of Balb/c mice after being implanted with MgG rods for different periods of time (0, 4 h, 12 h, 24 h, 72 h, and 7 days). n = 4 biologically independent animals in (a, c–d). n = 3 biologically independent animals in (g). A representative image of three biologically independent animals from each group is shown in (e). Data are presented as mean values ± SD.
